Numbers and tendencies and luck

Before I tell you anything my brain thinks about brackets, watch this video.

Then play this song while you read the rest of this post, because it always plays in my head when I fill out my bracket, like it’s a time-lapse scene in a movie when a computer crunches some numbers and then cuts to a close-up shot of a pencil writing a school name on a line.

It’s has to be the dramatic background music. Anyway, here are some things I’ve come to realize about brackets.

There are going to be upsets. Lots of them. If you don’t have a 12-seed beating a 5, you’re doing it wrong. Even having two 12′s advancing isn’t usually that bad of a strategy.

If you watched the video, you already know the following, but sometimes it helps to see it written down:

  • 11-seeds win one-third of the time in the first round.
  • 10-seeds win 40 percent of the time in the first round.
  • A 2-seed has been bounced before reaching the Sweet 16 in 14 of the last 15 years.
  • Not one, but two teams with double-digit seeds have made the Sweet 16 in 12 of the last 15 years.

I’ll show you my bracket before we talk strategy:

(Do me a favor and at least don’t insult me publicly for my picks.)

The numbers in the video are fairly convincing. You should pick, minimum, a 12, an 11 and at least one 10-seed in the first round, and then advance one or two of them to the Sweet 16 while knocking out a 2-seed. It seems like having a 10 beat a 2 kills two birds with one stone, but I decided not to because I couldn’t decide which one it should be (looking at you, Kansas) and didn’t want to risk blowing the points.

That’s where strategy comes in. Say you play the numbers at boot a 2-seed early. Don’t forget 2′s are pretty good teams. If you guess wrong and that team makes a run, you’re screwed in your pool. The same goes for any other upsets you pick.

You have to get over the idea of creating the perfect bracket. It’s not going to happen.

If you want to have a successful pool, the key is to play teams with high upset potential while minimizing foolish risk. You won’t win by playing it safe, but your risks have to be calculated.

Take Syracuse, for example. They aren’t going to win the tournament without Fab Melo, but they weren’t going to before anyway. But say you think Cuse is a train wreck without him and will lose early. Maybe you think the Orange won’t make it to the Sweet 16. But who do you put there instead?

Kansas State vs. Southern Mississippi is the 8-9 game that plays into Syracuse’s path. It was also one of the first round games (if you haven’t noticed, I’m refusing to acknowledge the NCAA refers to these as  the “second round,” claiming 60 teams get a bye from the play-in round) that gave me the most trouble when filling out my bracket.

Even if you decide you want Syracuse to lose at that point, you have to put K-State or Southern Miss there instead. If you guessed the wrong team in the opening (second) round, now that’s two games you aren’t getting any points for. Plus, Syracuse still has good players, and it has played without Melo for a little bit this year, so maybe it doesn’t lose after all.

Push the Orange ahead and don’t risk the points. If you pick Southern Miss to beat K-State and Cuse and it happens, you look like a genius, but if not, which is likely, you lose out on points and get behind.

The picture of my bracket was taken before the Melo news (I didn’t like Syracuse this year anyway), and I’m not so sure I like them to reach the Elite Eight anymore, but I’m also not sure I like Wisconsin going that far either. Decision, decisions.

Most years I look at the bracket and see at least one 12-seed right away that I think will beat a 5. This year, I don’t like any of them, but I have to pick one regardless. Long Beach State could be dangerous, but I like New Mexico’s chances, and the same can be said for Wichita State-VCU. I picked Harvard in the year of Lin, but I don’t have a lot of reason other than that. Especially with Syracuse down, it looks like Vanderbilt could get to the Elite 8, but now I’m just overthinking everything.

Three stats I look for in teams are scoring efficiency, three-point percentage and assist/turnover ratio. ESPN gives you the stats for points for and against, but I don’t pay much attention to that. That is a result of the style the team plays. If you run down and try to score as many points as possible (Iona), it will be higher, and if you have a slower pace (Wisconsin), it will be lower. Three-point rate is important for underdogs — if you want to pull of the upset, you have to be able to hit from distance. Assist/turnover ratio can alert you to teams that can’t take care of the basketball and/or are too one-dimensional.

Then there’s the Lin factor. There’s not really a reason for picking Harvard other than I needed a 12-seed to win. But there are so many games that will come down to the last minute, to one play, to whether or not a ball brushed someone’s leg before going out of bounds, that it’s impossible to think numbers can predict that type of stuff.

Stats can tell you generally what will happen, but other things like that are out of human reach. It’s just going to come down to luck, and picking a team because you like its mascot and know a girl there ends up being just as good of a reason as rebounding margin or out-of-conference RPI.

If I had the man power and the time, I’d create a stat for luck based on opponents’ free throw percentage and the percentage they actually shot against a given team, but who knows what that would show anyway. KenPom has a meter for luck, which it defines as “the deviation in winning percentage between a team’s actual record and [its] expected record using the correlated gaussian method,” another metric he created.

But I don’t so that’s why my bracket looks the way it does.

Other stuff you might be interested in:

  • Joining The Griffin’s bracket group on ESPN. I can’t give you any money, but the winner will get a write-up in the school paper and an interview from yours truly.
  • This cool map of all the NCAA teams. I like maps.
  • Knowing that Canisius played six games against tournament teams this year: two against Loyola and Iona, one against St. Bonaventure and one against UNLV. Lost ‘em all. The Griffs also played Lamar and Syracuse last year (beat Lamar).